DESCRIPTION (Applicant's Description) Cure rates in the acute lymphoblastic leukemias (ALL) now approach 75-80 percent in children and 30 percent in adults. These rates have reached a plateau and produce late adverse sequelae in many survivors. The hypothesis underlying this proposal is that improved outcomes will originate from laboratory investigation of the biology of ALL, leading to more specific therapies with less harmful effects of treatment. Our clinical trials result in an 82 percent 5-year event-free survival for ALL; yet, we remain unable to predict which children will relapse, making it difficult to tailor specific therapies for those at high risk. The challenges in adult ALL are even more daunting. We have a unique opportunity to test whether advances in the treatment of childhood ALL can be applied to a subset of adults whose lymphoblasts have similar pathogenetic features, and whether biologic differences can be exploited to develop more effective therapies. This Program combines focused and interactive cellular, molecular and clinical investigations, designed to develop the most effective, least toxic therapy for each patient. Project 5 will evaluate the prognostic significance of the TEL/AML1 gene, and its role in defining genetic events critical for the transformation of hematopoietic cells. This project will also develop an animal model of TEL/AML1 leukemia to provide insights into its molecular pathogenesis and a mechanism for testing novel therapies. In infants with MLL fusion genes, we will define interactions among cell death pathways at diagnosis and relapse (Project 6). In Project 7, we will define molecular pathways resulting in T-ALL through analysis of human leukemia cells, and develop genetic models of T-ALL in the zebrafish to dissect the functional implications of mutations that lead to the human disease, and identify genetic suppressors of the malignant phenotype. In close collaboration with investigators in the other projects, gene expression profiling using DNA microarrays and computer algorithms for pattern recognition will be used in Project 4 to seek molecular predictors of outcome for optimal assignment to specific therapies, and to identify new targets for antileukemic drug development. Molecular markers of residual leukemia will be applied to determine patients at risk for relapse (Project 9), and to develop novel therapeutic approaches based on interactions between host cells and tumor cells. We will also undertake trials of tumor vaccination and adoptive immunotherapy, and develop trials to modulate tumor-initiated angiogenesis (Project 10). Finally, we will integrate our findings into clinical trials to balance the quality of life against the rate of cure in both children and adults with ALL, thus maximizing the therapeutic index (Project 4). By understanding the pathogenesis and pathophysiology of ALL, and by translating laboratory findings made in the context of clinical trials, we seek to improve outcome for patients with ALL, which has historically served as a model and stimulated similar advances in other malignancies.